CN109061495A - A kind of hybrid energy-storing battery failure diagnostic method - Google Patents
A kind of hybrid energy-storing battery failure diagnostic method Download PDFInfo
- Publication number
- CN109061495A CN109061495A CN201810890449.8A CN201810890449A CN109061495A CN 109061495 A CN109061495 A CN 109061495A CN 201810890449 A CN201810890449 A CN 201810890449A CN 109061495 A CN109061495 A CN 109061495A
- Authority
- CN
- China
- Prior art keywords
- battery
- feature vector
- neural network
- network classifier
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Abstract
The present invention relates to battery failure diagnostic techniques in a kind of energy-storage system, especially a kind of hybrid energy-storing battery failure diagnostic method, to the voltage of single battery, electric current, the information signal datas such as temperature extract, and noise is removed by signal message data of the filtering algorithm to acquisition, obtain the sample of the faulty information of grandfather tape, extract the feature vector of battery condition from raw information by analysis again, and use this feature vector as the input signal of modified neural network classifier algorithm, fault feature vector and the one-to-one battery data training rules of fault type are established simultaneously, for training, test, the satisfactory diagnosis algorithm of diagnostic accuracy after after tested, for carrying out fault location to the battery module in actual motion.Advantage is: diagnostic method speed is fast, high-efficient, can be improved fail battery Detection accuracy, cuts off fail battery in time.
Description
Technical field
The present invention relates to battery failure diagnostic techniques in a kind of energy-storage system, especially a kind of hybrid energy-storing battery pack event
Hinder diagnostic method.
Background technique
In the power system, user demand side pipe reason can be effectively realized with energy storage technology, eliminates peak-valley difference round the clock,
Smooth load reduces power supply cost, while can promote the utilization of renewable energy, improves the stability of network system operation simultaneously
Grid power quality is improved, guarantees the reliability of power supply.As the battery system of the standby energy, whether battery operating status
Normally, normal, the reliable and safe operation of various equipment in application field is directly affected.
The health degree of battery pack depends on the single battery that health status is worst in battery pack.One battery pack is usually
It is composed in series by several single batteries or battery module, between each single battery in the battery pack being grouped after tested and preferably
There are still performance difference, these differences are in the During Process of Long-term Operation of battery because the minute differences of environment such as temperature difference can not
With degree generate new difference.By longtime running, Individual cells performance is decreased obviously, and seriously affects battery performance, very
To causing the accident;In addition, the performance decline of single battery and failure can reduce the state-of-charge value of battery pack.
Existing battery failures diagnostic techniques is broadly divided into two aspects: first is that the internal resistance or conductance using battery are held
Amount estimation and judgement;Second is that finding the correlation of battery performance failure and one or several parameter of battery, pass through real-time monitoring
Parameters variation compared between same batteries monomer battery carries out comprehensive descision.However the shortcoming of above two mode exists
In:
(1) when detecting to battery status, measuring device also can be different from measurement result caused by the difference of measurement method;
(2) (when high power charging-discharging, battery can generate temperature rise, will affect battery electricity for temperature and cell operating status when measuring
The activity of pole) difference also result in the inaccuracy of measurement result;
(3) difference of battery variety and capacity also results in the large error of measurement result.
Summary of the invention
A kind of accurate detection fail battery, in time is provided it is an object of the invention to shortcoming according to prior art
Cut off fail battery, the hybrid energy-storing battery failure diagnostic method that speed is fast, high-efficient.
Purpose of the present invention is realized by following approach:
A kind of hybrid energy-storing battery failure diagnostic method, is characterized by, includes the following steps:
1) voltage of single battery, the information signal data of electric current and temperature are extracted, and by filtering algorithm to the information of acquisition
Signal data removes noise, obtains the sample of the faulty information of grandfather tape,
2) a kind of follow-on neural network classifier is provided, extracts battery condition from the sample of the faulty information of grandfather tape
Feature vector, and use this feature vector as the input signal of neural network classifier algorithm:
1. by the feature vector construction feature vector sample database of the battery condition extracted, this feature vector sample database
In data sample include input feature value dimension and output failure mode label;
2. according to the input feature value dimension of data sample in feature vector sample database and output failure mode label
Neuronal quantity needed for neural network classifier is designed, the neuron includes fault feature vector and fault type;
3. each fault feature vector in neural network classifier is normalized, standardization, so that each failure
Feature vector is all in unified magnitude;
4. start the learning algorithm of neural network classifier, it, will treated fault signature according to the initial training of design rule
Vector and fault type establish mapping relations one by one, obtain learning training rule;
5. the learning algorithm by neural network classifier is trained the sample in feature vector sample database, supplement and
Learning training rule is modified, until neural network classifier reaches the nicety of grading requirement of design, acquisition test training rules;
6. neural network classifier is to the battery failures information sample except feature vector sample database according to test training rules
This carries out identification verifying;If the nicety of grading requirement of design cannot be reached, using this test training rules as learning training
5. regular return step carries out retraining;
7. this test training rules is as neural network classification if test training rules reach the nicety of grading requirement of design
Practical training rules in device;
8. carrying out failure to the battery module in actual motion using the neural network classifier for obtaining practical training rules to examine
Disconnected: the information signal data of single battery unifies the fault feature vector of magnitude by being converted into the battery pack in actual motion
Afterwards, corresponding fault type is mapped directly to according to practical training rules, establishes and is associated with and exports associated data.
In actual use, the fault message signal data that by practical training rules is not identified new when generation
When, neural network classifier will be again started up learning algorithm, and new fault message signal data is trained and is included in practical
In training rules, the training rules in neural network classifier are updated to adapt to a variety of different fault diagnosises.
The present invention provides a kind of hybrid energy-storing battery failure diagnostic methods, the advantage is that: using can constantly change
Into neural network classifier, by training, test, practical training rules are constantly updated in learning algorithm, so as to
Fail battery Detection accuracy is improved, cuts off fail battery in time, avoids the overall performance for influencing energy-storage battery group, diagnostic method
Speed is fast, high-efficient.
Detailed description of the invention
Fig. 1 is the step flow diagram of hybrid energy-storing battery failure diagnostic method of the present invention.
The present invention is described further below with reference to embodiment.
Specific embodiment
Most preferred embodiment:
Referring to attached drawing 1, the present invention is using improved artificial intelligence fault grader diagnostic method.To the electricity of single battery
The information signal datas such as pressure, electric current, temperature extract, and are made an uproar by signal message data removal of the filtering algorithm to acquisition
Sound obtains the sample of the faulty information of grandfather tape, then extracts the feature vector of battery condition from raw information by analysis,
And use this feature vector as the input signal of modified neural network classifier algorithm, while establishing fault feature vector and event
Hinder the one-to-one battery data training rules of type, for training, testing, the satisfactory diagnosis of rear diagnostic accuracy after tested
Algorithm, for carrying out fault location to the battery module in actual motion.
A kind of hybrid energy-storing battery failure diagnostic method of the present invention, includes the following steps:
1) voltage of single battery, the information signal data of electric current and temperature are extracted, and by filtering algorithm to the information of acquisition
Signal data removes noise, obtains the sample of the faulty information of grandfather tape,
2) a kind of follow-on neural network classifier is provided, extracts battery condition from the sample of the faulty information of grandfather tape
Feature vector, and use this feature vector as the input signal of neural network classifier algorithm:
1. by the feature vector construction feature vector sample database of the battery condition extracted, this feature vector sample database
In data sample include input feature value dimension and the output failure mode label fault type of battery (description);
2. according to the input feature value dimension of data sample in feature vector sample database and output failure mode label
Neuronal quantity needed for neural network classifier is designed, the neuron includes fault feature vector and fault type;
3. each fault feature vector in neural network classifier is normalized, standardization, so that each failure
Feature vector is all in unified magnitude;Such as to temperature samples Tx1Normalizing, standardization are carried out, then after normalizing standardizes
Temperature samples are (Tx1-Tmin)/(Tmax-Tmin), wherein TmaxFor the maximum temperature in temperature samples, TmaxFor in temperature samples
Minimum temperature;
4. start the learning algorithm of neural network classifier, it, will treated fault signature according to the initial training of design rule
Vector and fault type establish mapping relations one by one, obtain learning training rule;
5. the learning algorithm by neural network classifier is trained the sample in feature vector sample database, supplement instruction
Practice rule, until the nicety of grading that neural network classifier reaches design requires (training correct coincidence rate reach 95% or more),
Obtain test training rules;
6. neural network classifier is to the battery failures information sample except feature vector sample database according to test training rules
This carries out identification verifying (examining generalization ability);If the nicety of grading requirement of design cannot be reached, this is tested into training rule
Then retraining is 5. carried out as learning training rule return step;
7. this test training rules is as neural network classification if test training rules reach the nicety of grading requirement of design
Practical training rules in device;
8. carrying out failure to the battery module in actual motion using the neural network classifier for obtaining practical training rules to examine
Disconnected: the information signal data of single battery unifies the fault feature vector of magnitude by being converted into the battery pack in actual motion
Afterwards, corresponding fault type is mapped directly to according to practical training rules, establishes and is associated with and exports associated data.
In actual use, the fault message signal data that by practical training rules is not identified new when generation
When, neural network classifier will be again started up learning algorithm, and new fault message signal data is trained and is included in practical
In training rules, the training rules in neural network classifier are updated to adapt to a variety of different fault diagnosises.
The not described part of the present invention is same as the prior art.
Claims (1)
1. a kind of hybrid energy-storing battery failure diagnostic method, is characterized by, includes the following steps:
1) voltage of single battery, the information signal data of electric current and temperature are extracted, and by filtering algorithm to the information of acquisition
Signal data removes noise, obtains the sample of the faulty information of grandfather tape,
2) a kind of follow-on neural network classifier is provided, extracts battery condition from the sample of the faulty information of grandfather tape
Feature vector, and use this feature vector as the input signal of neural network classifier algorithm:
1. by the feature vector construction feature vector sample database of the battery condition extracted, this feature vector sample database
In data sample include input feature value dimension and output failure mode label;
2. according to the input feature value dimension of data sample in feature vector sample database and output failure mode label
Neuronal quantity needed for neural network classifier is designed, the neuron includes fault feature vector and fault type;
3. each fault feature vector in neural network classifier is normalized, standardization, so that each failure
Feature vector is all in unified magnitude;
4. start the learning algorithm of neural network classifier, it, will treated fault signature according to the initial training of design rule
Vector and fault type establish mapping relations one by one, obtain learning training rule;
5. the learning algorithm by neural network classifier is trained the sample in feature vector sample database, supplement and
Learning training rule is modified, until neural network classifier reaches the nicety of grading requirement of design, acquisition test training rules;
6. neural network classifier is to the battery failures information sample except feature vector sample database according to test training rules
This carries out identification verifying;If the nicety of grading requirement of design cannot be reached, using this test training rules as learning training
5. regular return step carries out retraining;
7. this test training rules is as neural network classification if test training rules reach the nicety of grading requirement of design
Practical training rules in device;
3) failure is carried out to the battery module in actual motion using the neural network classifier for obtaining practical training rules to examine
Disconnected: the information signal data of single battery unifies the fault feature vector of magnitude by being converted into the battery pack in actual motion
Afterwards, corresponding fault type is mapped directly to according to practical training rules, establishes and is associated with and exports associated data.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810890449.8A CN109061495A (en) | 2018-08-07 | 2018-08-07 | A kind of hybrid energy-storing battery failure diagnostic method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810890449.8A CN109061495A (en) | 2018-08-07 | 2018-08-07 | A kind of hybrid energy-storing battery failure diagnostic method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN109061495A true CN109061495A (en) | 2018-12-21 |
Family
ID=64832163
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810890449.8A Pending CN109061495A (en) | 2018-08-07 | 2018-08-07 | A kind of hybrid energy-storing battery failure diagnostic method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109061495A (en) |
Cited By (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110018425A (en) * | 2019-04-10 | 2019-07-16 | 北京理工大学 | A kind of power battery fault diagnosis method and system |
CN110133508A (en) * | 2019-04-24 | 2019-08-16 | 上海博强微电子有限公司 | The safe early warning method of electric automobile power battery |
CN110146817A (en) * | 2019-05-13 | 2019-08-20 | 上海博强微电子有限公司 | The diagnostic method of lithium battery failure |
CN110426637A (en) * | 2019-07-04 | 2019-11-08 | 佛山科学技术学院 | A kind of battery failures diagnostic method neural network based and device |
CN111901158A (en) * | 2020-07-14 | 2020-11-06 | 广东科徕尼智能科技有限公司 | Intelligent home distribution network fault data analysis method, equipment and storage medium |
CN112632850A (en) * | 2020-12-14 | 2021-04-09 | 华中科技大学 | Method and system for detecting abnormal battery in lithium battery pack |
CN112710956A (en) * | 2020-12-17 | 2021-04-27 | 四川虹微技术有限公司 | Battery management system fault detection system and method based on expert system |
CN113655391A (en) * | 2021-08-26 | 2021-11-16 | 江苏慧智能源工程技术创新研究院有限公司 | Energy storage power station battery fault diagnosis method based on LightGBM model |
CN114323691A (en) * | 2021-12-28 | 2022-04-12 | 中国科学院工程热物理研究所 | Gas circuit fault diagnosis device and method for compressed air energy storage system |
WO2022149824A1 (en) * | 2021-01-08 | 2022-07-14 | 주식회사 엘지에너지솔루션 | Apparatus and method for managing battery |
CN114781551A (en) * | 2022-06-16 | 2022-07-22 | 北京理工大学 | Battery multi-fault intelligent classification and identification method based on big data |
CN116879788A (en) * | 2023-07-04 | 2023-10-13 | 广东鸿昊升能源科技有限公司 | Energy storage electric cabinet safety detection method, device, equipment and storage medium |
Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101634605A (en) * | 2009-04-10 | 2010-01-27 | 北京工业大学 | Intelligent gearbox fault diagnosis method based on mixed inference and neural network |
CN104714175A (en) * | 2013-12-12 | 2015-06-17 | 北京有色金属研究总院 | Battery system fault diagnosis method and system |
CN107422266A (en) * | 2017-03-15 | 2017-12-01 | 中国电力科学研究院 | A kind of method for diagnosing faults and device of high capacity cell energy-storage system |
-
2018
- 2018-08-07 CN CN201810890449.8A patent/CN109061495A/en active Pending
Patent Citations (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101634605A (en) * | 2009-04-10 | 2010-01-27 | 北京工业大学 | Intelligent gearbox fault diagnosis method based on mixed inference and neural network |
CN104714175A (en) * | 2013-12-12 | 2015-06-17 | 北京有色金属研究总院 | Battery system fault diagnosis method and system |
CN107422266A (en) * | 2017-03-15 | 2017-12-01 | 中国电力科学研究院 | A kind of method for diagnosing faults and device of high capacity cell energy-storage system |
Non-Patent Citations (1)
Title |
---|
张洪瑾: "基于模糊神经网络的掘进机液压系统故障诊断研究", 《中国优秀硕士学位论文全文数据库 工程科技II辑》 * |
Cited By (16)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110018425A (en) * | 2019-04-10 | 2019-07-16 | 北京理工大学 | A kind of power battery fault diagnosis method and system |
CN110133508B (en) * | 2019-04-24 | 2022-04-01 | 上海博强微电子有限公司 | Safety early warning method for power battery of electric automobile |
CN110133508A (en) * | 2019-04-24 | 2019-08-16 | 上海博强微电子有限公司 | The safe early warning method of electric automobile power battery |
CN110146817A (en) * | 2019-05-13 | 2019-08-20 | 上海博强微电子有限公司 | The diagnostic method of lithium battery failure |
CN110426637A (en) * | 2019-07-04 | 2019-11-08 | 佛山科学技术学院 | A kind of battery failures diagnostic method neural network based and device |
CN111901158A (en) * | 2020-07-14 | 2020-11-06 | 广东科徕尼智能科技有限公司 | Intelligent home distribution network fault data analysis method, equipment and storage medium |
CN112632850A (en) * | 2020-12-14 | 2021-04-09 | 华中科技大学 | Method and system for detecting abnormal battery in lithium battery pack |
CN112710956A (en) * | 2020-12-17 | 2021-04-27 | 四川虹微技术有限公司 | Battery management system fault detection system and method based on expert system |
CN112710956B (en) * | 2020-12-17 | 2023-08-04 | 四川虹微技术有限公司 | Expert system-based battery management system fault detection system and method |
WO2022149824A1 (en) * | 2021-01-08 | 2022-07-14 | 주식회사 엘지에너지솔루션 | Apparatus and method for managing battery |
CN113655391A (en) * | 2021-08-26 | 2021-11-16 | 江苏慧智能源工程技术创新研究院有限公司 | Energy storage power station battery fault diagnosis method based on LightGBM model |
CN114323691A (en) * | 2021-12-28 | 2022-04-12 | 中国科学院工程热物理研究所 | Gas circuit fault diagnosis device and method for compressed air energy storage system |
CN114323691B (en) * | 2021-12-28 | 2023-06-23 | 中国科学院工程热物理研究所 | Air path fault diagnosis device and method for compressed air energy storage system |
CN114781551A (en) * | 2022-06-16 | 2022-07-22 | 北京理工大学 | Battery multi-fault intelligent classification and identification method based on big data |
CN114781551B (en) * | 2022-06-16 | 2022-11-29 | 北京理工大学 | Battery multi-fault intelligent classification and identification method based on big data |
CN116879788A (en) * | 2023-07-04 | 2023-10-13 | 广东鸿昊升能源科技有限公司 | Energy storage electric cabinet safety detection method, device, equipment and storage medium |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109061495A (en) | A kind of hybrid energy-storing battery failure diagnostic method | |
KR102543921B1 (en) | Method for estimating battery life for a mobile device based on relaxation voltages | |
CN112731159B (en) | Method for pre-judging and positioning battery faults of battery compartment of energy storage power station | |
CN109856299A (en) | A kind of transformer online monitoring differentiation threshold value dynamic setting method, system | |
CN110133536A (en) | Determine system, the method and apparatus of the index of battery group object internal leakage electric current | |
CN113933732A (en) | New energy automobile power battery health state analysis method, system and storage medium | |
Li et al. | A weighted Pearson correlation coefficient based multi-fault comprehensive diagnosis for battery circuits | |
CN114152825B (en) | Transformer fault diagnosis method and device and transformer fault diagnosis system | |
WO2023024851A1 (en) | Battery equalization method and system | |
CN115480180A (en) | New energy battery health diagnosis and analysis method | |
CN109615273A (en) | A kind of electric car electrically-charging equipment method for evaluating state and system | |
CN108693478A (en) | A kind of method for detecting leakage of lithium-ion-power cell | |
CN115877205A (en) | Intelligent fault diagnosis system and method for servo motor | |
CN111707943A (en) | Battery simulation-based electric vehicle charging fault early warning method and system | |
CN115616428A (en) | Charging-detecting integrated electric vehicle battery state detection and evaluation method | |
CN114325433A (en) | Lithium ion battery fault detection method and system based on electrochemical impedance spectrum test | |
CN109752664A (en) | A kind of charging detects the method and application of battery core internal resistance in battery pack | |
CN116572747B (en) | Battery fault detection method, device, computer equipment and storage medium | |
CN109270383A (en) | A kind of non-intrusion type charging pile automatic testing method | |
CN116298947B (en) | Storage battery nuclear capacity monitoring device | |
CN117491872A (en) | Reconfigurable battery module fault multistage diagnosis method | |
CN113391214A (en) | Battery micro-fault diagnosis method based on battery charging voltage ranking change | |
He et al. | How long will my phone battery last? | |
Tao et al. | Lithium-ion Battery Performance Degradation Recognition Method based on SOC Estimation | |
CN117054892B (en) | Evaluation method, device and management method for battery state of energy storage power station |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181221 |
|
RJ01 | Rejection of invention patent application after publication |